________________
MAY, 1892.)
THE INSCRIPTIONS OF PIYADASI.
153
this point of view, anomalies like ustana and anususţi can be easily explained. ustana is only another way of spelling ufhána. The cerebralization, for which anuśásti supplies no pretext, could creep into the pronunciation of anusatthi under the influence of the analogy which it suggests with forms like sillha, anusit tha. A practice of this kind, extending even to words in which it has no etymological justification, is certainly not without example in Hindi usage. I content myself with quoting the use of the groups gr, tr in Jain Prakrit,to used to represent merely a doubled 'g or t, and that even when it is not justified by otymology, - in pôgralá, i. e. pôggala (pudgall), as well as in udagra. Prof. Weber has not on this account dreamed of suggesting that the pronunciation ugra, wagra has been preserved, but very rightly concludes that we must everywhere read gga.
The preceding remarks do not exhaust the instances in which wo are permitted to infor that the orthography of the Edicts of Piyadasi is not strictly phonetic. Other spellings deserve, from this point of view, to be noticed here. Some are significant by their very character and by their inconsistencies; others, either better preserved or altered more than the mean level of phonetic decay permits, reveal in turn either an accidental imitation of the cultivated idiom, or the contemporary existence of a popular language into which the mode of writing of our inscriptions artificially introduces a regularity unkuown in practice,
In the first category is contained the use of t. This brings me again to Dr. Pischel's remarks. I must confess that I can no longer hold to the opinion, originally expressed by me, tbat the sign tat Khálsi was only another form of +. I admit that this sign, literally kya, corresponds to a special shade of pronunciation, although it does not appear to be easy to define it. The concurrence of the forms kalingya, kaliingyésu, kalingy dni, which Dr. Bühler has been the first to identify at Khâlsi (XIII, 5, 6,), does not throw much light on the problem ; but to whatever conclusion we are led, it will remain none the less certain that the engravers have displayed a singular inconsistency. According to Dr. Pischel himself, beside seventeen instances in which the suffix ika is written ikya, there are seven in which the spelling ika is retained. It is very clear that one or other of these two methods of writing does not accord with the exact pronunciation. What are wo to say about the Dehli inscriptions, in which we find £ in two isolated examples, in ambavydikd and qdhakósiķani (Col. Ed. VII-VIII, 2), whereas everywhere else the suflix invariably retains the form ikea ?
I confess that I find some difficulty in avoiding an explanation, which, at the first glance, will appear singular and rash. In various coins of Spalngadama, of Spalirisos (Sallet, p. 154), and of Gondophares (p. 169), we find dharmiasa side by side with the ordinary form dharmikasa. On the other hand, the coins of Lysias (ibid, p: 154) have alternately lisikasa and lisiasa. The pronunciations ika and iya do not appear to have belonged to the same period of phonetic development, and it is tempting to conclude that the popular pronunciation was iya, (or ia, which is the same thing), of which ila represents the learned spelling; that, in fact, people read the latter iya, as seems to be proved by the writing lisikasa for lisiyasa. The sign t ought hence to be considered as a compromise between the real pronunciation, indicated by the y, and the tatsama orthography represented by the k. The spelling alikasadala must be explained by some play of etymology, which, in order to lend to the foreign name a Hindi appearance, seems to have sought in the first portion of the word for the Prikrit alika, aliya, correspond ing to the Sanskțit alika, I do not underrate the difficulties of this solution. If it were certain, it would lend a singular confirmation to my method of considering the orthography of our inscriptions, but I recognise that it is in no way certain. I only put it forward as a conjecture, which is, in my opinion, a likely one, and I do not propose to take advantage of it elsewhere for any more general conclusions. If we neglect it, and content ourselves with a simple statement of the facts, we find at all events that, at least in this particular point, the
1. cf. Weber. Bhagavatt, pp. 387 and ff.
